Artificial intelligence can predict flood disasters

In the future, artificial intelligence will help predict flood disasters much more precisely. Google has been working on this early warning system since 2017.

Since 2000, the number of global flood disasters has more than doubled, increasing by 134 percent. This causes damage amounting to around 50 billion US dollars every year.

Google now wants to remedy this and provide people in these high-risk areas with an improved early warning system. The US company has been working on the project since 2017, in which artificial intelligence is intended to make the prediction of flood disasters more precise.

Google wants to use artificial intelligence to predict flood disasters

Around 1.5 billion people worldwide are at constant risk of flooding. That's almost a fifth of the world's population. For this reason, Google wants to improve early warning systems and thus “save thousands of lives every year”.

In the future, people in risk areas should receive precise information sooner and therefore in a timely manner. Using Google's real-time flood forecasting system, the group can provide alerts via Google Search, Maps, Android notifications and Flood Hub.

Using machine learning to predict floods

In the specialist magazine Nature Google has now presented its research progress. In the article, the search engine giant shows how machine learning technologies can “significantly improve” flood forecasting. Google is particularly interested in countries “where little data on flooding is available”.

Using these AI-based technologies, we have increased the reliability of currently available global forecasts from zero to five days on average and improved forecasts for regions in Africa and Asia to be comparable to those currently available in Europe.

Real-time flow forecasts are also possible thanks to the technologies used. According to Google, Flood Hub can deliver data up to seven days in advance and cover rivers in over 80 countries.

Research on the use of artificial intelligence to predict flood disasters

To research the FloodHub tool, Google worked with scientists, governments, international organizations and NGOs. The first pilot early warning system was launched in the Ganges-Brahmaputra river basin in India in 2018.

Google's hydrological models can now predict river flooding. To do this, they use publicly available weather data and physical information about catchment areas.

If certain data is missing, artificial intelligence comes into play to predict flood disasters. Then, with the help of machine learning, Google can “train a single model on all available flow data.” In this way, “models can be trained worldwide and predictions can be made for every river location”.

Also interesting: